From Clone to Cloud in 60 Seconds: One-Click LLM Council Deployment

Published: December 2025

Prerequisites

Before deploying, you'll need:

• An OpenRouter account with API key
• A Railway or Render account (free to create)
• For local development: Docker Desktop installed

The Problem

You want to try LLM Council. You read the documentation, you understand the 3-stage deliberation process. Now you want to run it.

The traditional path looks like this:

git clone https://github.com/amiable-dev/llm-council.git
cd llm-council
pip install -e ".[http]"
export OPENROUTER_API_KEY=sk-or-v1-...
llm-council serve

Not bad for developers. But what about:

• n8n users who want to integrate with workflow automation?
• Evaluators who want to test before committing to a tech stack?
• Teams who need a shared council endpoint without everyone managing API keys?

For these users, deployment friction is the #1 barrier to adoption.

The Solution: One-Click Deploy

With ADR-038 (our Architectural Decision Record for deployment strategy), we've added one-click deployment to two platforms:

Platform	Deploy	Cost
Railway		~$5/mo
Render		Free tier

Click the button, select the repository, add your API keys (OPENROUTER_API_KEY and LLM_COUNCIL_API_TOKEN). Done.

Why Two Platforms?

Different users have different needs:

Railway: For Production

Railway is our recommended platform for production use:

• No cold starts - Paid Railway instances stay warm (unlike free tier platforms)
• Webhook reliability - Critical for n8n/Make/Zapier integration
• Template marketplace - Organic discovery among 2M+ developers
• Revenue sharing - Railway shares template revenue with creators, supporting open-source sustainability

Render: For Evaluation

Render's free tier is perfect for evaluation:

• 750 free hours/month - Plenty for testing
• Quick setup - No credit card required
• Blueprint support - Infrastructure-as-code

Caveat: Render Free tier spins down after 15 minutes of inactivity. This causes 30-60 second cold starts that will likely timeout webhooks (n8n/Zapier default timeout is often 30 seconds). For workflow automation integration, use Railway or Render paid tier.

Security First: API Token Authentication

A public council endpoint without authentication is a credit-draining vulnerability. With ADR-038, we've added LLM_COUNCIL_API_TOKEN authentication:

# Set in your deployment platform's environment
LLM_COUNCIL_API_TOKEN=your-secure-token-here

All protected endpoints now require a Bearer token:

curl -X POST https://your-app.railway.app/v1/council/run \
  -H "Authorization: Bearer your-secure-token-here" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "What is the best way to learn programming?"}'

The /health endpoint remains public for load balancer health checks.

Local Development with Docker Compose

For local testing and development:

# Clone the repo
git clone https://github.com/amiable-dev/llm-council.git
cd llm-council

# Create .env file
echo "OPENROUTER_API_KEY=sk-or-v1-..." > .env
echo "LLM_COUNCIL_API_TOKEN=my-local-token" >> .env

# Start the server
docker compose up --build

The council is now running at http://localhost:8000.

Step-by-Step: Railway Deployment

1. Click Deploy on Railway

1. Connect GitHub (if prompted)

2. The template deploys from amiable-dev/llm-council automatically

3. Connect your GitHub account if Railway prompts you

4. Configure Environment Variables

Railway will prompt you for:

Variable	Description
OPENROUTER_API_KEY	Your OpenRouter API key
LLM_COUNCIL_API_TOKEN	A secure token for API auth

Generate a secure token:

openssl rand -hex 16

1. Deploy

Railway builds and deploys automatically. Within 2-3 minutes, you'll have a live URL.

1. Test Your Deployment

# Health check (no auth required)
curl https://your-app.railway.app/health
# Expected: {"status":"ok","service":"llm-council-local"}

# API request (auth required)
curl -X POST https://your-app.railway.app/v1/council/run \
  -H "Authorization: Bearer your-token" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "What is the capital of France?"}'

A successful response returns the 3-stage deliberation results plus metadata:

{
  "stage1": [{"model": "...", "content": "..."}],
  "stage2": [{"model": "...", "evaluation": "...", "parsed_ranking": [...]}],
  "stage3": {"synthesis": "Paris is the capital of France..."},
  "metadata": {
    "aggregate_rankings": {"Response A": 1.5, "Response B": 2.0},
    "config": {"council_size": 3, "verdict_type": "synthesis"}
  }
}

Troubleshooting

Symptom	Cause	Solution
401 Unauthorized	Missing or incorrect token	Verify Authorization: Bearer <token> header format
500 Internal Server Error	Invalid OpenRouter API key	Check OPENROUTER_API_KEY is set correctly
502 Bad Gateway	App crashed on startup	Check deployment logs for errors
Timeout on first request	Cold start (Render free tier)	Wait 60s and retry, or use Railway
Health check fails	Port binding issue	Ensure app binds to $PORT (handled automatically)

n8n Integration

With a deployed council endpoint, n8n integration is straightforward. In your n8n workflow, add an HTTP Request node with these settings:

Setting	Value
Method	POST
URL	https://your-app.railway.app/v1/council/run
Authentication	Header Auth
Header Name	Authorization
Header Value	Bearer your-token-here
Body	JSON

Request Body:

{
  "prompt": "{{$json.question}}",
  "verdict_type": "binary"
}

Secure Token Storage

Store your API token in n8n environment variables (Settings → Variables) as COUNCIL_TOKEN, then reference it as {{$env.COUNCIL_TOKEN}} in the Header Value field.

Input Field

The {{$json.question}} syntax assumes the previous node outputs a JSON field named question. Adjust to match your workflow trigger.

See the full n8n integration guide for complete workflow examples including code review automation and support ticket triage.

Revenue Sustainability

Why Railway as the primary platform? Beyond technical fit, Railway's Open Source Kickback program provides revenue sharing for template creators. This creates a sustainability path for open-source projects.

By using our official Railway template, a portion of your Railway spend supports LLM Council development.

What's Next?

• Deployment Guide - Detailed platform-specific guides
• n8n Integration - Workflow automation setup
• HTTP API Reference - Full API documentation
• ADR-038 - Technical decision record

---

This post implements ADR-038: One-Click Deployment Strategy.

Explore the source code: github.com/amiable-dev/llm-council